Ground Point Cloud Detection Using Segmented Distance Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for detecting ground point cloud points in autonomous vehicles often result in false obstacle detection, as the ground is mistakenly identified as an obstacle, leading to reduced accuracy in obstacle detection.
Innovation Solution
A method and apparatus that determine a segmentation plane and segment the point cloud into sub-clouds based on distance thresholds, with closer point clouds having lower thresholds and sparser point clouds having higher thresholds, to accurately identify ground point cloud points.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single distance threshold is used to determine ground point cloud points, then the detection process is simple, but the detection accuracy is reduced due to varying point cloud densities at different distances
Solution Approach 1:
The patent applies local quality by using different distance thresholds for different spatial regions. Specifically, it divides the point cloud into multiple ranges based on distance from the lidar: a first range (0.5m-2m) uses a first threshold, a second range (2m-5m) uses a second threshold, and a third range (5m-10m) uses a third threshold. This allows the detection method to adapt to the varying point cloud densities at different distances, improving detection accuracy without requiring a completely complex system overhaul.
Solution Approach 2:
The patent changes the threshold parameter based on the distance range. By adjusting the distance threshold parameter according to different spatial regions (using smaller thresholds for closer points and larger thresholds for farther points), the method optimizes ground point detection accuracy for each region while maintaining a relatively simple overall detection framework.
2Reliability
If a uniform threshold is applied to all point clouds, then the detection method is easy to implement, but false obstacle detection increases
Solution Approach 1:
The patent implements local quality by applying different distance thresholds to different spatial regions. The method divides point clouds into three distance ranges and applies progressively larger thresholds for farther ranges, which reduces false obstacle detection while maintaining ease of implementation through a structured, rule-based approach rather than complex algorithms.
Solution Approach 2:
The patent segments the point cloud data into multiple distance ranges (first range: 0.5m-2m, second range: 2m-5m, third range: 5m-10m) and applies different thresholds to each segment. This segmentation strategy improves reliability by reducing false detections while keeping the implementation straightforward through clear range-based classification.
3Measurement precision
If distance-based segmentation is used to improve detection accuracy, then detection precision improves, but processing complexity increases
Solution Approach 1:
The patent segments the point cloud into three distance-based groups (0.5m-2m, 2m-5m, 5m-10m) and applies different thresholds to each segment. This segmentation improves detection precision by accounting for varying point densities while maintaining relatively simple processing through straightforward range classification and threshold application.
Solution Approach 2:
The patent changes the threshold parameter based on distance ranges, using smaller thresholds for closer point clouds and larger thresholds for farther point clouds. This parameter adaptation improves detection precision while keeping processing complexity manageable through a systematic, distance-based parameter selection approach.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
Embodiments of the present disclosure relate to a method and apparatus for detecting ground point cloud points. The method may include: determining a segmentation plane and a ground based on a point cloud collected by a lidar; segmenting the point cloud into a first sub point cloud and a second sub point cloud based on the segmentation plane, where distances between point cloud points of the first sub point cloud and the lidar are shorter than a distance between the segmentation plane and the lidar, and distances between point cloud points of the second sub point cloud and the lidar are not shorter than the distance between the segmentation plane and the lidar; and determining the point cloud points whose distances from the ground are smaller than a first distance threshold in the first sub point cloud as ground point cloud points, and determining the point cloud points whose distances from the ground are smaller than a second distance threshold in the second sub point cloud as the ground point cloud points, where the first distance threshold is smaller than the second distance threshold. This embodiment improves the accuracy of detection of the ground point cloud points.